Mycobacterium tuberculosis is currently the etiologic agent that is responsible for the most deaths worldwide. Mycobacterium avium is the most common cause of systemic bacterial infection in American AIDS patients and, despite the public health importance of these diseases, the mechanisms of virulence remain unknown. We have adopted three strategies to identify mycobacterial virulence factors: 1) We are attempting to complement the virulence defect in a virulent M. intracellular strain with genes from a virulent M. avium strain. These genetic complementation studies should allow us to identify M. avium genes which contribute to virulence. 2) We are establishing an "ivet"-like system for mycobacteria with the ivet system genes that are expressed in vivo and thought to be important for virulence or persistence. A necessary element for the ivet technology is a mycobacterial auxotroph. We have identified the mycobacterial trylophan synthetase gene and are presently trying to generate an auxotroph based on the sequence data. 3) The catalose-peroxidose (katG) protein has been shown to be a virulence factor in M. bovis BCG. We are currently evaluating the role of KatG in the virulence of M. Tuberculosis using genetic complementation.